Soccer Refereeing Comprehensive Evaluation on SoccerRefBench (Overall)
66.56AccuracySoccerRef-Agents
Evaluation Results
| Method | Links | |
|---|---|---|
| SoccerRef-AgentsModel Category=Agentic System2026.04 | 66.56 | |
| GPT-4oModel Category=Commercial APIs2026.04 | 64.58 | |
| Gemini 2.5 FlashModel Category=Commercial APIs2026.04 | 55.17 | |
| Claude 4.5 SonnetModel Category=Commercial APIs2026.04 | 55.12 | |
| Qwen3-VL-32BModel Category=Open-Source Models2026.04 | 45.93 | |
| Qwen3-VL-8BModel Category=Open-Source Models2026.04 | 39.16 |